Models for the Cyclic Resistance of Silts and Evaluation of Cyclic Failure during Subduction Zone Earthquakes

Abstract

This report describes several advances in the cyclic failure assessment of silt soils with immediate and practical benefit to the geotechnical earthquake engineering profession. First, a database of cyclic loading test data is assembled, evaluated, and used to assess trends in the curvature of the CRR-N (cyclic resistance ratio - the number of equivalent cycles) relationship. This effort culminated in a plasticity index-dependent function which can be used to estimate the exponent b in the power law describing cyclic resistance, and may be used to estimate the cyclic resistance of silt soils as well as the number of equivalent loading cycles anticipated for subduction zone earthquakes. Statistical models for the cyclic resistance ratio and cyclic strength ratio are presented in this report. The SHANSEP (Stress History and Normalized Soil Engineering Properties)-inspired functional form of these models have been trained and tested against independent datasets and finalized using a combined dataset to provide reasonable estimates of resistance based on the available data. These models can be used to provide provisional estimates of the CRR-N and cyclic strength ratio power laws for cyclic shear strain failure criteria ranging from 1 to 10%, within certain stated limitations. The ground motion records within the NGA Subduction Project which have been released to the public to-date are implemented to examine the role of subduction zone earthquake characteristics on the number of equivalent loading cycles for a wide range of soils with exponents b ranging from 0.05 (moderate plasticity silt and clay) to 0.35 (dense sand). This analysis shows that the number of loading cycles for a given magnitude subduction zone earthquake is larger than those previously computed, whereas the corresponding magnitude scaling factors for use with the Simplified Method span a smaller range as a result of the ground motion characteristics. Owing to the large variability in the computed equivalent number of loading cycles, consideration of the uncertainty is emphasized in forward analyses. The work described herein may be used to estimate cyclic resistance of intact non-plastic and plastic silt soils and corresponding factor of safety against cyclic failure for a range in cyclic shear strain failure criteria, to plan cyclic laboratory testing programs, and to calibrate models for use in site response and nonlinear deformation analyses in the absence of site-specific cyclic test data. As with any empirical approach, the models presented herein should be revised when additional, high-quality cyclic testing data become available.

PEER Workshop on Liquefaction Susceptibility

“…That the hysteretic behavior of these transitional soil specimens can evolve throughout loading highlights the role of earthquake duration (i.e., N) on the potential for exhibiting sand-like behavior. Thus, short duration crustal earthquakes may not produce sufficient loading cycles to trigger sand-like behavior, whereas longer duration (e.g., subduction zone) earthquakes, which can produce greater than 100 cycles of loading, depending on the power law exponent b describing the CRR-N relationship Stuedlein et al 2021), can lead to the transient loss of strength. Furthermore, it was observed that specimens that exceeded  = 3% in the first cycle (i.e., subjected to large CSR) often required a number of additional cycles to satisfactorily establish the ultimate hysteretic behavior, indicating that significantly larger shear strains than those associated with common cyclic failure criteria are necessary to make determinations of sandlike or clay-like behavior.…”